Advertise
MCP.so
Sign In

Meridian MCP Server

@meridianmcp

Repository

Basic information

Category

Developer Tools

Transports

stdio

Author

meridianmcp

Submitted by

Adam Camerer

About Meridian MCP Server

Shared memory and coordination for AI coding sessions. Persistent sprint board, human-in-the-loop queue, and parallel session protection. Works with Claude Code, Cursor, Codex, and any MCP client.Shared memory and coordination for AI coding sessions. Persistent sprint board, huma

#developer-tools#ai#coding#memory#coordination#mcp

Meridian MCP Server config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "meridian": {
      "type": "http",
      "url": "https://usemeridian.us/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Meridian MCP Server tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

Fetch the live tool list directly from this server's endpoint using the button above.

Overview

What is Meridian?

Meridian is an open-source MCP server that gives AI coding sessions persistent shared memory. It provides a shared task log, pinned decisions, a human-in-the-loop queue, and tiered handoffs so sessions don't lose context. It works with Claude Code, Cursor, Cline, Claude Desktop, or any MCP client.

How to use Meridian?

Clone the repository, run the install script, and start the server (default dashboard at http://localhost:7878). Wire it into your AI client by adding a .mcp.json config (for hosted, use an API key; for self-host, use a stdio command). Auto‑checkpoint hooks can also be installed to inject context on session start and snapshot on stop.

Key features of Meridian

  • Persistent session memory (shared task log, pinned decisions, goal state)
  • 50+ MCP tools (start_session, log_task, claim_task, set_decision, generate_handoff, etc.)
  • Symbol‑level parallel safety – two sessions can safely edit the same file at class/function granularity
  • Live work queue – planners inject sprint items; executors pick them up mid‑run
  • HITL recommended option – flag a safe default; dashboard highlights it for one‑key selection
  • Tiered handoffs (L0/L1/L2) – a fresh session resumes in seconds
  • Webhook intake – push events from LangGraph / Autogen / custom agents
  • Dashboard with sessions, tasks, sprint board, swimlane timeline, and pinned decisions

Use cases of Meridian

  • Eliminate context debt when coding across multiple AI sessions (no more re‑explaining architecture or constraints)
  • Team coordination – point multiple local instances at a shared Postgres for instant shared sessions
  • Auto‑checkpoint coding workflows – install hooks so every session start injects project context and every stop writes a delta handoff
  • Planning chat via claude.ai web – bridge all 50+ tools into the browser with the dnakov/claude‑mcp extension

Frequently asked questions

1

What problem does Meridian solve?

Every AI coding session boots blind. Meridian provides shared memory so sessions see the same task log, pinned decisions, and goal state. When context fills up, a new session resumes from a compressed handoff in seconds.

2

How do I install and run Meridian?

Clone the repository and run the install script (`./install.sh` on macOS/Linux, `.\install.ps1` on Windows), then start with `pixi run start`. A dashboard opens at `http://localhost:7878`. Data persists in `./data/meridian.db` by default. No cloud required for local use.

3

Where is session data stored?

By default, state lives in a local SQLite file (`./data/meridian.db`). You can use a shared Postgres database by setting the `MERIDIAN_DB_URL` environment variable. For the hosted tier, data is stored on Neon Postgres.

4

What MCP clients are supported?

Meridian works with Claude Code, Cursor, Cline, Claude Desktop, and any standard MCP client. It can be configured via stdio (self‑host) or HTTP (hosted). The claude.ai web client is supported via a Chrome extension bridge.

5

How does authentication work?

For the hosted tier, you generate an API key (`sk_meridian_...`) in your settings and pass it in the `Authorization` header. Self‑hosted instances run locally with no authentication by default. OAuth and email magic link are available for the hosted version.

Comments

More Developer Tools MCP servers

LastPing

LastPing

Know when your AI agents, cron jobs and CI pipelines fail, stall or go quiet. Your agent sets up monitors from a plain description, tells you what is broken and why, reads run history and traces, and writes notes on inci

Mcp-Warden

ermoluk

Local-first credential broker for MCP. AI coding agents (Claude Code, Cursor, Codex…) call APIs, SSH and databases through scoped presets; secrets stay in your OS keychain and never enter the agent's context.

4

ZeroBeacon — 1000-Tool Collision-Proof Commerce Router

DavidFox998

ZeroBeacon.ai — Collision-proof commerce router for AI agents. 1000 MCP tools: commerce, trust, billing, sovereign ID + everyday tools. LIVE.

offprompt

offprompt

MCP server that lets Claude Code, Cursor or Codex ask you for API keys on a local page and write them to .env. The agent only ever sees key names.

2

TranscriptFetch MCP Server

TranscriptFetch

Model Context Protocol (MCP) server for TranscriptFetch: fetch YouTube transcripts, search, channels, and playlists from any MCP client.

Advertise